Across the world, a quiet revolution is underway in how work gets done. Recent surveys show nearly nine in ten organisations now use some form of AI in their operations, and the use of generative AI tools has more than doubled in just a couple of years.

Employers expect these technologies to reshape their businesses before 2030.

The financial returns on automation look attractive at first glance. Studies by major consultancies report that around three-quarters of organisations say their AI and automation investments meet or exceed expectations.

Cup pressure peaks as Shabana host Kenya Police in blockbuster quarter-final Football

Upper Hill coach Walgwe unperturbed despite drawing giants in Prescott Cup group phase Rugby

In some areas, such as robotic process automation, first-year returns can be very high, and long-term gains even higher. A handful of big companies have saved billions by using AI for fraud detection, logistics and other back-office tasks.

However, these headline stories hide a more sobering reality. Many boards expect new technology to pay for itself within a year, yet research shows typical AI projects now take two to four years to break even.

Only a small minority recoup their investment in under twelve months. For Kenyan and African firms operating under high borrowing costs and thin margins, that mismatch between expectations and reality can be fatal.

When returns are slower than promised, leaders cut budgets, sour on the technology and blame AI, rather than their own rushed approach.

Even in rich countries, most companies are still experimenting rather than transforming. One leading report estimates that more than half of today’s work activities could already be automated with existing tools.

Yet only a tiny fraction of business leaders describe their AI rollouts as “mature”, and fewer than one in five firms have fully modernised their processes around AI.

For Africa, this gap between potential and practice is even wider. Kenya is rightly seen as a digital pioneer, but the continent as a whole still accounts for a very small share of the global AI market.

Infrastructure gaps, patchy data, and limited research funding make it harder to build systems that understand African languages, markets and realities. If we simply import foreign tools trained on foreign data, we risk automating bias, misunderstanding local customers and entrenching dependence on overseas vendors.

The labour implications are equally complex. Globally, forecasts suggest tens of millions of jobs will be created and destroyed by 2030, with a net gain but a high level of churn as roles change.

Entry-level work is particularly exposed: a significant share of junior, routine and support tasks can be automated. For a young country like Kenya, where hundreds of thousands enter the labour market each year, this should ring alarm bells.

At the same time, the right kind of automation can boost productivity and open new opportunities. Studies show workers become about one-third more productive when they use generative AI tools, and programmers complete tasks dramatically faster with AI assistants.

In African contexts, this could mean doctors seeing more patients, teachers preparing better lessons, small businesses doing proper bookkeeping, and government officers serving citizens quicker—if they have access to tools, connectivity and training.

The real problem is not that AI fails, but that organisations misuse it. Research from MIT, IDC and others shows that most enterprise generative AI pilots fail to produce measurable financial benefits within six months, and a large majority of proofs-of-concept never make it into production.

Globally, only about a third of technology projects are completed on time, on budget and with all goals met.

For Kenya, this should be a cautionary tale as the government rolls out a national AI strategy and businesses rush to deploy chatbots, copilots and automated decision systems.

The lesson from global experience is clear: technology alone does not deliver efficiency. What matters is whether organisations redesign their processes, invest in data and skills, and build strong governance around how AI is used.

Policy makers should therefore treat AI infrastructure, digital skills and worker protection as long-term public investments, not short-term gadgets.

Business leaders need to stop chasing fashionable tools and instead pick a few high-impact use cases tightly linked to national priorities—such as reducing post-harvest losses, improving hospital workflows or sealing leakages in public finance—and commit to them for several years.

Workers and unions must be brought into these conversations from the start, with clear plans for upskilling and transitions.

If Kenya and its African peers get this right, automation can help turn our demographic boom into an economic dividend rather than a crisis. If we get it wrong, we risk ending up as low-wage data providers and test beds for systems designed elsewhere, while the real value—and the most productive jobs—flow out of the continent.

Ann Kinyanjui, CPA, is a finance and business leader with extensive experience managing multi-country operations, driving strategic growth, and improving financial performance across diverse industries.